Do States’ Immigrant-Friendly Policies Improve the Health of Children of Immigrants? The Impact of Driver’s License Policies for Undocumented Immigrants and “Sanctuary” Policies on Access and Use of Health Care

Heather Koball and Seth Hartig

National Center for Children in PovertyBank Street Graduate School of Educa on

This work was generously supported by the Robert Wood Johnson Foundation (RWJF), Evidence for Action Program. Any opinions expressed are those of the authors alone and should not be construed as representing the opinions of RWJF. The paper authors would like to offer a special thanks to Dr. James Kirby, senior researcher in the Center for Financing, Access, and Cost Trends at the Agency for Healthcare Research and Quality, for his very helpful assistance in the data analysis completed for this report.

1

Introduction

This study examines the health impacts of local and state immigration-related policies that support undocumented immigrants and their families. States and localities have adopted two types of “immigrant-friendly” policies: (1) allowing undocumented immigrants to obtain driver’s licenses and (2) limiting state/local law enforcement’s involvement in federal immigration enforcement efforts (i.e., sanctuary policies). In the past years, the federal government has increased pressure on states/localities to participate in federal immigration enforcement, through threats of reduced federal funding (Chau and Morse, 2012) and immigration raids that specifically target “sanctuary” places (Ortiz, 2017).

Assessing the impact of these policies on health is particularly salient because of the current COVID-19 pandemic. If the 10.5 million undocumented immigrants living in the United States are unable or afraid to access health care, this could worsen the pandemic, as well as threaten their lives. Recent studies have illustrated the negative health outcomes for undocumented immigrants and their families, including their U.S. citizen children, when state/local law enforcement vigorously participate in federal immigration enforcement efforts. However, we know of no study to date that has examined the health impacts of “immigrant-friendly” policies, such as local/state sanctuary policies and state driver’s licenses policies.

The research question we sought to answer in this study was “Do sanctuary policies and driver’s licenses for undocumented immigrants increase the chances that children of these immigrants will receive adequate medical and dental care?” We focused on access to preventive health care, including whether children in immigrant families (1) have a usual source of care (USC) provider, (2) have unmet medical needs, (3) visited the dentist in the past year, or (4) had recent well-child visits. We merged policy data from a database of state-level immigrant policies with data from the Medical Expenditure Panel Survey (MEPS), and examined variation in immigrant children’s health outcomes in different state policy contexts.

Our analyses demonstrated that sanctuary policies and driver’s licenses for undocumented immigrants improved preventive health outcomes among children of immigrants. At varying levels of significance, these policies were shown to increase the likelihood that children, especially Latino children, living with non-citizen, immigrant parents have a usual source-of-care (USC) provider, and that they significantly reduce the likelihood that children in these households have unmet medical needs. The study also demonstrated that these policies significantly increase the likelihood that Latino children living with non-citizen, immigrant parents receive preventive dental care. This effect held regardless of whether the children were U.S. citizens.

These results are important given the wide variation in state responses to recent upswings in federal immigration enforcement and the importance of access to health care in addressing an epidemic such as COVID-19. We hope that a clearer understanding of the health impacts of driver’s license and sanctuary policies can help inform immigration policy decisions at the state and local level.

Background

About a quarter of U.S. children are children of immigrants (Urban Institute, 2019). An estimated 5.6 million children in the United States, or about 7 percent, live with an undocumented immigrant parent who lacks permission to live or work in the United States (Passel and Cohn, 2018). Of these children, over 75 percent are U.S.-born citizens, living in “mixed-status” families (Passel and Cohn, 2018).

Emerging evidence shows that having undocumented immigrant parents affects various dimensions of children’s well-being and health, likely due to the lack of resources and the ever-present anxiety about the future that

2

undocumented immigrants face. Having undocumented parents had been tied to disruptive housing arrangements when parents are faced with deportation (Amuedo-Dorantes and Arenas-Arroyo 2019), greater anxiety and depression (Dreby 2012; S. R. Potochnick and Perreira 2010; Gulba et al 2015), and fewer years of educational attainment by early adulthood (Bean et al, 2011). Children of immigrants have low health insurance rates, even when they are U.S. citizens, but especially when they have undocumented immigrant parents (Kaiser Family Foundation, 2019; Weathers et al, 2008; Acevedo-Garcia and Stone, 2008; Ziol-Guest and Kalil, 2012).

However, the effects of growing up with undocumented parents may vary across the United States. State and local governments have varied in their responses to increasing federal immigration enforcement, creating different immigration policy contexts state by state and locality by locality.

The relationship between driver’s licenses and immigrant family health

Until 1993, undocumented immigrants could legally obtain a driver’s license in all states. California was the first state to restrict driver’s licenses for undocumented immigrants, followed by Arizona in 1996. Other states quickly followed suit and by 2011, only three states (Utah, New Mexico, and Washington) allowed undocumented immigrants to obtain a driver’s license. However, a subset of states began to change course beginning around 2012 to allow the provision of licenses and, as of the time this report was written, 15 states, the District of Columbia, and Puerto Rico allow driver’s licenses for undocumented immigrants. Because of the 2005 REAL ID federal law requirements, these licenses are distinct from the ones issued to citizens or legal immigrants.

Prior qualitative work on health and human service needs of children of undocumented immigrants suggests that lack of transportation resulting from state restrictions on driver’s licenses for undocumented immigrants creates a significant barrier to receiving health care (Koball, Capps, Hooker, Perreira, Campetella, Pedroza, Monson, and Huerta, 2015). Interviews with undocumented youth reveal high rates of fear of driving in states that bar undocumented immigrants from obtaining driver’s licenses, particularly given the presence of police checkpoints intended to catch people driving without a license (Roth 2014). In places with stricter immigration enforcement, a traffic stop for undocumented immigrants may be the first step toward detention and deportation (Waslin 2013). This is compounded by immigration enforcement activities that have resulted in undocumented immigrants being arrested for driving without a license, which may lead to deportation (Capps, Rosenblum, Chishti, and Rodriguez, 2011). Lack of driver’s licenses mean more limited mobility, with strong implications for ensuring that members of undocumented families are able to easily get to doctor’s appointments or hospital care. As far as the authors know, there has been no published study that has provided a quantitative analysis of the health impacts of driver’s license policies.

Sanctuary policies, state and local non-participation in federal immigration enforcement

The degree to which states restrict participation in federal immigration enforcement primarily revolves around state and local responses to the federal Secure Communities program. Secure Communities was launched as a pilot in 2008 under President Bush and expanded nationally during President Obama’s first term. Under Secure Communities, law enforcement agencies submit fingerprints of arrestees for checks against Department of Homeland Security (DHS) databases. If an immigration violation is found, Immigration and Customs Enforcement (ICE) officials may issue a detainer request to local law enforcement to hold an individual for up to 48 hours so that ICE can take custody. Submitting fingerprints to DHS is required by law; detaining people for ICE to pick up is not required by law.

3

Starting in 2011, some states, cities, and counties passed policies to limit cooperation with ICE detainer requests. These policies have colloquially been referred to as “sanctuary policies.” Some state and local law enforcement agencies have been concerned about violating individuals fourth amendment rights, as some courts have ruled ICE detainer policies do. Some local law enforcement agencies are concerned about violating trust between local law enforcement and immigrant communities, which could hinder their ability to address criminal activity in their localities. Other state/local agencies were concerned about using limited local/state law enforcement resources to address the federal priority of immigration enforcement. Research showed that Secure Communities led to deportations of many noncitizens with minor or no criminal convictions.

In late 2014, Secure Communities was replaced with the Priority Enforcement Program (PEP) (Chishti and Hipsman, 2015), which focused on detaining and deporting immigrants who were national security threats, had been convicted of serious crimes, or were new arrivals in the U.S. PEP also allowed local law enforcement agencies to set further limits on cooperation with federal immigration enforcement (Rosenblum, 2015). President Trump’s January 2017 executive order on interior immigration enforcement ended PEP, reinstated the Secure Communities program, and reestablished expanded enforcement priorities (including the detention/deportation of undocumented immigrants without criminal records).

Some localities have proactively participated in federal immigration enforcement through voluntary participation in the 287g program, a program that deputizes local law enforcement to act as federal immigration enforcement agents. Some states and localities have passed their own laws that aim to limit immigration or to make life more difficult for immigrants in those communities. These proactive laws are described in greater detail in the appendix; however, this analysis focuses on the effects of sanctuary laws on health outcomes.

Effects of immigration enforcement on health of immigrants

The psychological effects of enforcement policies have been well-documented. In a 2008 national survey, 72 percent of foreign-born Latinos reported worrying a lot about being deported or having a family member deported (Lopez and Minushkin 2008). In addition to the studies above, interviews with immigrant families and children in Ohio and New Jersey revealed children’s strong fears of parents’ deportation, likewise revealing the strong reverberation of enforcement efforts throughout communities (Dreby 2012). A survey study conducted by Hatzenbuehler et al (2015) found that Latinos living in states with such policies had higher rates of poor mental health than Latinos living in states with less restrictive immigration policies. Several studies show also that state immigration enforcement efforts —287(g) agreements, use of the E-Verify employment verification system, and passage of state policies restricting immigrants’ rights—likely drive undocumented immigrants out of high-enforcement states toward lower-enforcement states (Bohn, Lofstrom, and Raphael 2013; Leerkes, Leach, and Bachmeier 2012; Parrado 2012; Watson 2013).

Surveys of undocumented immigrants have revealed high rates of fear of accessing public benefits due to fears of deportation and fear of accessing medical services because of immigration status (Amuedo-Dorantes, Puttitanun, and Martinez-Donate 2013; Berk and Schur 2001). Hacker et al (2012) conducted a qualitative study in which they found health care providers reported negative impacts of ICE activities, such as raids, detention, and deportation on immigrant patients’ physical and psychological health. Wang and Kaushal (2018) found that Secure Communities arrangements increased the proportion of Latino immigrants with mental health distress, that Section 287(g) Task Force agreements worsened mental health distress, and that Section 287(g) jail enforcement agreements worsened the health of Latino immigrants.

4

Policies around public health insurance for immigrants (1996-present)

In assessing the impact of driver’s licenses and sanctuary policies on health outcomes, it is important to consider other state policy variation that could disparately influence health outcomes among immigrant families, especially policies regarding eligibility for public health benefits like Medicaid. After controlling for income, immigrant families have lower rates of take-up of public benefits than families headed by citizens, due in large part to immigrants’ restricted eligibility for these programs (Ku and Bruen 2013). The Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA, known commonly as “welfare reform”) excluded most lawful permanent residents (LPRs) from accessing most federally-funded public benefits, including Medicaid, during their first five years with LPR status. Undocumented immigrants have always been ineligible for Medicaid; however, states can choose to expand public health insurance to these otherwise ineligible immigrants using state funds. States decide whether to use state funds to allow federally excluded groups to participate in state public health insurance programs (Fix 2009; Kretsedemas and Aparicio 2004; Zimmerman and Tumlin 1999), leading to variation in immigrants’ eligibility for public benefits by state.

Since 2009, states have had the option to use federal funds to provide Medicaid/CHIP benefits to certain specific groups, including lawfully present children and pregnant women, even during the five-year federal ban, and to provide coverage for prenatal care for pregnant women, regardless of immigration status. The timing of the expansion of eligibility and benefits for immigrants also differs across states. Currently, seven states allow all legal immigrants to access public health insurance within the five-year waiting period for federal coverage, 31 states allow legal immigrant children to access public health insurance within the waiting period, six states allow undocumented immigrant children to access public health insurance, and 32 states allow immigrant pregnant women to access public health insurance within the waiting period. For more information on public health policies for immigrants, please go to https://www.urban.org/features/state-immigration-policy-resource

Stricter enforcement has been shown to reduce enrollment in public benefits, regardless of a child’s eligibility, for both children of undocumented immigrants and children of legal immigrants (Vargas and Pirog, 2016). Local 287(g) agreements in effect between 2004 and 2009 have been shown to increase the risk of food insecurity for households that are low-income, Mexican, and non-citizen that contain children (Potochnick, Chen, & Perreira, 2016; Potochnick et al, 2017) Qualitative studies have suggested that fear of accessing health services or public benefits has been shown to be heightened after immigration enforcement raids (Chaudry, Capps, Pedroza, Castaneda, Santos, and Scott, 2010). Watson (2010, 2014) and Vargas (2015) find that higher immigration enforcement deters Medicaid enrollment among children who are citizens by their mothers who are non-citizens and immigrants.

Impact on greater population health

Beyond the impact state immigration-related policies have on children of immigrants’ health, there are two mechanisms through which immigrants’ lack of access to primary health care could harm the broader population health. The first is through infectious diseases. If immigrant populations have low vaccination rates or deviate from recommended vaccination schedules because of lack of access to a usual source of care, this can put the broader community at risk for outbreaks. In particular, children who are not insured or whose parents fear taking them to the doctor might not receive regular well-child visits, resulting in missed or delayed vaccinations (Buelow and Van Hook, 2008). A second mechanism is through the utilization of scarce health resources (Steege, Baron, Davis, Torres-Kilgore and Sweeney, 2009). Specifically, immigrants without a usual primary health care provider might resort to using scarce ED resources when their health problems become severe (Jacobs, Tovar, Hung, Kim, Ye, Chiang, and Goldfrank, 2002). Greater access to primary care might minimize their utilization of these resources.

5

Our contributions

We are not aware of any previous studies that have analyzed the impact of state driver’s license policies toward immigrants on adverse health and health care outcomes among immigrant families, or of studies that focus on limited cooperation policies to support these health outcomes. To fill this gap, we examined several indicators of preventive care access, namely, whether the child: (1) has a usual care provider, (2) has unmet medical need, (3) had a check-up or well-child visit in the last year, (4) saw a dentist in the last 6 months or the past year. We utilized a database of state immigration enforcement and public benefits policies toward immigrants for all 50 states, covering each year from 2000 to 2016. Developed as part of a previous study (Gelatt, Koball, Bernstein, Runes, and Pratt, 2017), this database is publicly available for other researchers to use. https://www.urban.org/features/state-immigration-policy-resource

Methods

Our analyses draw on two primary cross-sectional data sources: 17 panels of the Medical Expenditure Panel Survey (MEPS) that includes in-depth data during the same time period on preventive health care receipt and immigration status, and linked the data by state of residence an original database of state policies toward immigrants between 2000 and 2016.

Data

Database of Immigrant-Related and Public Health Insurance Policies. Members of the research team had already created a state database covering 2000-2016 state enforcement policies and public health insurance (Medicaid/CHIP) policies in all 50 states and D.C. The database is publicly available at http://www.urban.org/features/state-immigration-policy-resource. This database was created by culling and standardizing information from a range of 50-state data sources, primarily provided by the National Immigration Law Center (NILC), the National Conference of State Legislatures (NCSL), Urban Institute Welfare Rules Database, and other sources. The standardized data were reviewed by experts in state immigration policy from NILC, NCSL, and the Urban Institute.

For this analysis, we use the policies on limited cooperation with Secure Communities (i.e., sanctuary policies), driver’s licenses, and public health insurance policies in the 50-state immigration policy database to examine their impact on the receipt of preventive health care among children of immigrants. This follows the approach successfully employed in a previous analysis of the impact of state immigration enforcement on the material hardship experienced by children of immigrants (Gelatt et al, 2017).

Medical Expenditure Panel Survey (MEPS) Data. To test our hypotheses, we required individual-level survey data that includes: (1) measures of preventive health care received by children, (2) state identifiers to link with the state policies, and (3) immigrant characteristics over the appropriate time period. We selected restricted-access MEPS data for our analysis because it is one of the few data sets that meet these criteria. MEPS includes a nationally representative panel of approximately 30,000 respondents annually and is a nationally representative survey for the U.S. non-institutionalized population collected by the Agency for Healthcare Research and Quality since 1996. Families are interviewed five times over two years. Interviews are conducted in English or Spanish, with interpretation for speakers of other languages. MEPS data contain a range of demographic, economic, and health outcomes. We included 2000-2016 MEPS data to capture events before and after some states changed their policies regarding limited cooperation and driver’s licenses for undocumented immigrants in the early to mid-2000s. We merged the MEPS data by year and state with the state immigrant policy database for the analyses.

6

Sample. We used the same coding categories for immigrant households similarly to how we have coded children and families in our prior analyses that examine the impact of state policies on children of immigrants. We used three categories of children to compare outcomes: (1) Child and Parents Native-Born: children living in households where both they and their parents are born in the U.S., which we used as a comparison group; (2) Parent Immigrant, Child Native-Born: children who were born in the U.S. and live in households in which at least one parent was born outside the U.S. and is not a naturalized U.S. citizen; and (3) Child and Parents, Immigrants: children living in households in which both they and their parents are born outside the United States and neither is a naturalized U.S. citizen.

We also performed the analyses with a subsample of only Latino children. The MEPS data, like most surveys, does not ask respondents whether they are legally residing in the United States. About 42 percent of non-U.S. citizen immigrants are undocumented (Passel 2019). Among Latino immigrants, about 65 percent of non-citizen, immigrants are undocumented (Baker 2019, Passel and Cohn 2018). We anticipate that the driver’s license and limited cooperation policies would have a stronger impact on the outcomes of children with undocumented immigrant parents. Because of this imprecision in our analysis, we anticipate that our results are the lower bound of the impact of these policies.

Measures drawn from MEPS - outcome variables:

Outcome variables. Our outcome variables include whether the child: (1) has a usual care provider, (2) had a well child visit or check up in the last year (3) had an unmet medical need in the past year or (4) saw a dentist last year, described in more detail below. These questions are asked of children more than once over the two-year observation period. We used the outcome data collected one year after the baseline survey to capture receipt of preventive care in the previous year.

• Usual source of care (USC) provider: The MEPS instrument asks individuals whether there is a specific person or place to which they would go if ill or in need health advice. Those who answer affirmatively were coded as having a USC provider.

• Unmet Need: The MEPS instrument first asks respondents whether they or a doctor thought they needed care, tests or treatment in the last 12 months. Individuals who report need are then asked whether they were able to get the care they needed and whether they were delayed in getting the care they needed. We use a dichotomous variable to identify children who either did not get needed care or who were delayed in getting care. Analogous measures were constructed for dental and prescription drug treatments. We further categorize Unmet Need by whether that need is specifically medically related.

• Visits: At each interview round, respondents report all medical events that occurred since the previous interview, or since the beginning of the year if it was the first interview. For office visits, individuals are asked to report the “main reason” for the visit. Among the possible responses are “checkup” and “well-child visit” and we identified individuals who had at least one such visit during the year. Measures incorporated into outcome variables utilizing visit data include whether children in the sample had a dental checkup in the last year, a dental checkup in the last 6 months, and whether children had a well-child visit or checkup in the last year.

State policy database - explanatory variables

In order to test how sanctuary policies and driver’s licenses affect health outcomes we merged the MEPS data with our database of state policies using MEPS’s state identifiers. This longitudinal database includes measures of local and state laws that restrict cooperation with ICE detainers issued through Secure Communities, driver’s

7

licenses, and state public benefits policies. We include policy measures for 2000-2016, to match the years in which outcomes are measured in the MEPS.

Driver’s License Policies: Whether or not a state had adopted a policy of allowing undocumented immigrants to legally obtain driver’s licenses.

Limited Cooperation Policies: This variable indicates whether localities within the state had policies to restrict cooperation ICE detainer requests. If some or all of the largest immigrant counties had such a policy within the state or the whole state had such a policy, the state was coded as having such a policy. More information about these policies is available at https://www.urban.org/features/state-immigration-policy-resource

Control variables:

We used data collected at baseline to measure a range of demographic variables, including the measures described below. We use state fixed effects to control for differences across states that do not change during the period of the analysis. We also controlled for state-level factors that may be correlated with state adoption of immigration enforcement policies, and that might disproportionately affect immigrant households’ receipt of health care, including other immigration enforcement policies as derived from the state immigration policy database described above.

On the family level, control variables included:

• Age of oldest parent (in years)• Age of child• Gender of child• Health insurance coverage of child• Family income as percentage of the poverty line, broken down into five income bands • The highest degree obtained by either parent (no high school, high school diploma or GED, some college,

4 year college degree or greater), • Interview language (English, Spanish, Other), • Number of children <18 in the household, and at the county level, • Number of doctors per 1000 residents, • Whether the family lived in an urban, suburban, or rural area (a trichotomous variable indicating

whether family lived in metropolitan area, a nonmetropolitan area adjacent to a metropolitan area, and a nonmetropolitan area non-adjacent to a metropolitan area)

• Region of residence (broken down into 9 MEPS regions comprised of New England, Middle Atlantic, East North Central, West North Central, South Atlantic, East South Central, West South Central, Mountain, and Pacific

• The parents rating of the child’s health status, as captured by dichotomous variables measuring health as “poor” to “excellent” on a five-point Likert-scale. Also included is a series of dichotomous variables indicating whether an individual has the following chronic conditions: angina, asthma, coronary heart disease, diabetes, emphysema/COPD, high blood pressure, myocardial infarction, other heart disease, and stroke

At the state level, tied to families by linking our state policy database with MEPS data, control variables included:• Percent Latino• Whether the state provides public health insurance to legal immigrants prior to the five-year bar• Whether the state provides public health insurance to undocumented immigrant children

8

• Whether the state provided TANF cash assistance to legal permanent residents who have not satisfied the five-year ban on receiving federal assistance

• Whether the state provided food assistance to legal permanent residents who have not satisfied the five-year ban on receiving federal assistance

• Whether the largest immigrant counties in the state have 287(g) agreements, through a task force model or a jail model

Demographic statistics

Table [1] shows weighted estimates of the characteristics of families in our sample, by family immigration status. Table [2] shows weighted estimates of the characteristics of Latino families in our sample, by family immigration status; as explained below, there is reason to expect that Latino families are especially impacted by the policies examined in this analysis.

9

10

11

Analytic methods

To address research question 1 of whether state immigration enforcement policies affect receipt of preventive care among children of immigrants, we used difference-in-difference analysis for each of the outcome variables. The difference-in-difference analysis was of the following form:

Yist = α + χDRLCst-1 + φIMMIGRANTist-1 + β1DRLCst-1*IMMIGRANTist-1 + φXist-1 + ηs + υt-1 + εist-1

where Y is the measure of receipt of preventive care, for child i in a low-income household in state s at time t, DRLCst-1 is an indicator for whether the state has a policy of allowing driver’s licenses to undocumented immigrants or has a limited cooperation policy at time t-1, IMMIGRANTist-1 indicates child i lives in a household with at least one immigrant parent at time t-1. We run the same model with IMMIGRANTist-1 defined as a child i who is an immigrant and who lives with at least one immigrant parent. The comparison group for both groups are U.S.-born children living with only U.S.-born parents. DRLCst-1*IMMIGRANTist-1 is an interaction term, and Xist-1 is a vector of the control variables described above. State and time fixed effects are ηs and υt-1, respectively, and the error term is εist-1. The models are estimated using ordinary least squares with standard errors clustered at the state level, in keeping with the standard approach for estimating difference-in-difference models.

The parameters of interest are β1, which captures the effect of being in a state that has implemented policies, controlling for all the covariates in the model. A positive coefficient on β1 for an outcome, such as having a usual source of care (USC) provider, would indicate that children of immigrants in states with driver’s licenses for undocumented immigrants are more likely to have a USC, relative to children of immigrants in states without licenses (this is compared to the difference in this outcome across these states for children of U.S.-born parents), providing suggestive evidence that these policies support preventive healthcare use among children of immigrants.

The fundamental identifying assumption of this type of difference-in-difference approach is that the relative difference in the outcomes measured (in this case, outcomes related to preventive health care use) between the treatment group (children living with immigrant parents) and the comparison group (children of U.S.-born parents), is similar in the set of states that have adopted the intervention being analyzed (driver’s license policies and/or limited cooperation policies) and separately in the set of states that have not after accounting for the control variables in the model. While this assumption is fundamentally untestable, we performed tests for the validity of this assumption along with robustness checks, described in more detail at the end of the paper.

Findings

Tables 3-6 provide the results of our difference-in-difference analysis using four models to explore the impact of these policies on our health outcomes, specifically the presence of a Usual Source of Care Provider (USC), Unmet Medical Need, Dental Checkup in the Last Six Months, and Any Well-Child Visits or Checkups in the Last Year. The tables detail the results of the analysis using four models. The first two models report on the results of the above analysis for all racial and ethnic groups (“all”), and the second two models report on the results when the population of the study is limited to children identified as Hispanic/Latino (“Latino”), the racial/ethnic group likely most impacted by these policies, most likely targeted by anti-immigration policies, and a group that previous studies (mentioned above) have indicated has experienced negative health and socioeconomic outcomes as a result of such policies. Within the “all” and “Latino” models, we used a model that examined the impacts of driver’s license policies alone (“Driver’s License model”), and another model that analyzed the impact of limited cooperation policies (“Limited cooperation model”).

12

13

The analyses above confirm the hypotheses that driver’s license policies and limited cooperation policies improve preventive health outcomes among immigrant populations, especially Latino immigrants, compared to U.S.-born children. Table [4] and [5] show significant increases in dental checkups among children who are U.S. born with immigrant parent(s) and an increase in immigrant children having a USC related to driver’s license policies and limited cooperation policies.

The impacts of these policies are stronger among Latino children. Table [6] demonstrates that among U.S.-born children with immigrant parents, driver’s license policies reduce unmet medical need (at a 5% level of significance). Among immigrant children with immigrant parents, driver’s license policies increase the likelihood of immigrant children with immigrant parents having a USC provider and increases the likelihood of having been to a well child visit or checkup in the last year, at a 5% level of significance. Table [7] shows that limited cooperation policies increase the likelihood of having a dental checkup in the past six months, and that such policies increase the likelihood of immigrant children with immigrant parents having a USC provider.

To test whether income affected our results, we ran separate models restricting the sample to low-income households, defined as households with incomes of less than 200% of the poverty line. We found similar outcomes for the low-income group.

As described above, concluding that state policies affect access to and use of preventive health care rests on the assumption that there were no substantial unobservable differences related to health care between children in immigrant households in the two types of states analyzed (those that adopted driver’s license and/or limited cooperation policies), relative to children in non-immigrant households in those two types of states. Our difference-in-difference analysis is an attempt to address this assumption because it compares the relative rates of preventive health care among children in immigrant households in these two types of states to those of children in non-immigrant households in those states. Provided that any differences between children in immigrant households and U.S.-born children on unobserved characteristics are consistent across these two types of states, significant findings suggest that state policies caused the differences in receipt of preventive health care.

Study Limitations•  Legal status of immigrants. The MEPS data, like most surveys, does not include information about non-

citizen immigrants’ legal status. We found that the results became stronger when we narrowed immigrant households to those who were most likely to be undocumented, suggesting that our results are a lower bound of the estimates.

• Selective out-migration. Some evidence suggests that immigration enforcement (or immigrant protections) pushes undocumented immigrants to friendlier states (Leerkes and Bachmeier, 2013; Lofstrom, Bohn, and Raphael, 2011; and Watson, 2013), but has not established whether these movers differ from immigrants who remain in high-enforcement states. It is possible that selective out-migration leads to immigrant families with more resources to move from states without these policies to states that have them. This could lead to a correlation of these policies with more access to preventive health care.

• Analysis of local policy variations. Although policy changes regarding driver’s licenses occur at the state level, law enforcement cooperation with immigration enforcement agencies is primarily a matter dictated by local governments and police forces. Local policy variations can be especially important when political environments at the local level - where limited cooperation agreements can be implemented—are more conducive to supporting undocumented immigrant families than are policies at the state level.

14

References

Acevedo-Garcia, D. & and Stone, L.C. (2008), “State Variation in Health Insurance Coverage for U.S. Citizen Children of Immigrants,” Health Affairs (Project Hope) 27, no. 2 (April 2008): 434–46, doi:10.1377/hlthaff.27.2.434. Amuedo-Dorantes and Arenas-Arroyo (2019)Amuedo-Dorantes, Catalina, Thitima Puttitanun, and Ana P. Martinez-Donate. 2013. “How Do Tougher Immigration Measures Affect Unauthorized Immigrants?” Demography 50 (3): 1067–91. doi:10.1007/s13524-013- 0200-x.Baker, Bryan. 2019. “Estimate of Lawful Permanent Resident Population in the United States: January 2015.” Washington, DC: Office of Immigration Statistics, Department of Homeland Security. https://www.dhs.gov/ sites/default/files/publications/lpr_population_estimates_january_2015.pdfBean, Frank D., Mark A. Leach, Susan K. Brown, James D. Bachmeier, and John R Hipp. 2011. “The Educational Legacy of Unauthorized Migration: Comparisons Across U.S.-Immigrant Groups in How Parents’ Status Affects Their Offspring1.” International Migration Review 45 (2): 348–385. doi:10.1111/j.1747- 7379.2011.00851.x.Berk, Marc L., and Claudia L. Schur. 2001. “The Effect of Fear on Access to Care Among Undocumented Latino Immigrants.” Journal of Immigrant Health 3 (3): 151–56. doi:10.1023/A:1011389105821.Bernstein, Hamutal, and Julia Gelatt. 2014. “Contexts of Immigrant Integration and Service Delivery.” In Annual Meeting. Albuquerque, NM.Buelow, V., & Van Hook, J. (2008) “Timely Immunization Series Completion among Children of Immigrants.” Journal of Immigrant & Minority Health, 10(1), 37–44. Bohn, Sarah, Magnus Lofstrom, and Steven Raphael. 2013. “Did the 2007 Legal Arizona Workers Act Reduce the State’s Unauthorized Immigrant Population?” Review of Economics and Statistics 96 (2): 258–69. doi:10.1162/REST_a_00429.Capps, R.; Marc R. Rosenblum, Muzaffar Chishti, and Christina Rodriguez. (2011). “Delegation and Divergence: 287(g) State and Local Immigration Enforcement.” Washington, DC: Migration Policy Institute. http://www. migrationpolicy.org/research/delegation-and-divergence-287g-state-and-local-immigration-enforcement.Carrasquillo, O.; Carrasquillo, A.;, & Shea, S. (2000). “Health Insurance Coverage of Immigrants Living in the United States.” American Journal of Public Health, 90(6), 917-923; Chau L. and Morse. A. (2012). “U.S. Supreme Court Rules on Arizona’s Immigration Enforcement Law.” Washington, DC: National Conference of State Legislatures.Chaudry, A; Randolph Capps, Juan Pedroza, Rosa Maria Castaneda, Rob Santos, and Molly M. Scott. 2010. “Facing Our Future: Children in the Aftermath of Immigration Enforcement.” Washington, DC: The Urban Institute.Chishti, M.; & Hipsman, F. (2015). “Sanctuary Cities Come Under Scrutiny, As Does Federal-Local Immigration Relationship.” 2015. Policy Beat. Washington, DC: Migration Policy InstituteDepartment of Health and Human Services (1998). Office of the Secretary. Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA); Interpretation of ‘‘Federal Public Benefit.”’ Federal Register / Vol. 63, No. 14 /Tuesday, August 4, 1998. Retrieved from https://www.govinfo.gov/content/ pkg/FR-1998-08-04/pdf/98-20491.pdf. Derose, K.; Bahney, B., Lurie, N.; & Escarce, J. (2009). “Review: Immigrants and Health Care Access, Quality, and Cost.” Medical Care Research & Review, 66(4), 355–408. Dreby, J. (2012). “The Burden of Deportation on Children in Mexican Immigrant Families.” Journal of Marriage and Family 74 (4): 829–45. doi:10.1111/j.1741-3737.2012.00989.x.Dyssegaard Kallick and Roldan (2017). Expanding Access to Driver’s Licenses Getting a License Without Regard to Immigration Status. Retrieved from Fiscal Policy Institute Website: http://fiscalpolicy.org/wp-content/ uploads/2017/01/FINAL-Drivers-licenses-report-2017.pdfFix, Michael E. 2009. Immigrants and Welfare: The Impact of Welfare Reform on America’s Newcomers. Russell Sage Foundation.

15

Gelatt J., Koball, H.; Bernstein, H.; Runes, C.; & Pratt, E. (2017). State Immigration Enforcement Policies: How They Impact Low-Income Households. Retrieved from http://www.urban.org/sites/default/files/ publication/90091/state-immigration-enforcement-policies.pdf.Hacker et al (2012)Hatzenbuehler et al (2015)Johnston, G.; & Morse, A. (2011). “2010 Immigration-Related Laws and Resolutions in the States.” Washington, DC: National Conference of State Legislators.Jacobs, D.; Tovar, J.; Hung, O.; Kim, M.; Ye, P.; Chiang, W.; & Goldfrank, L. (2002). “Behavioral Risk Factor and Preventive Health Care Practice Survey of Immigrants in the Emergency Department.” Academic Emergency Medicine, 9(6), 599–608.Johnson, B. (2019, January 7). Murphy and top N.J. Dems support drivers licenses for undocumented immigrants. But advocates fear it’s being stalled. NJ.com, Retrieved from https://www.nj.com.Kaiser Family Foundation, “Health Coverage of Immigrants,” Washington, DC, 2019. Retrieved from https://www. kff.org/disparities-policy/fact-sheet/health-coverage-of-immigrants/. Koball, H.; Randy Capps, Sarah Hooker, Krista Perreira, Andrea Campetella, Juan Manuel Pedroza, William Monson, and Sandra Huerta, (2015).“Health and Social Service Needs of U.S.-Citizen Children with Detained or Deported Parents” Washington, DC: The Urban Institute.Kretsedemas, Philip, and Ana Aparicio, eds. 2004. Immigrants, Welfare Reform, and the Poverty of Policy. Westport, CT: Praeger.Ku, L. (2009). “Health insurance coverage and Medical Expenditures of Immigrants and Native-born Citizens in the United States.” American Journal of Public Health, 2009, 99(7), 1322–1328; Ku, L.; & Bruen, B. (2013). “Poor Immigrants Use Public Benefits at a Lower Rate than Poor Native-Born Citizens.” Washington, DC: Cato Institute. http://social.cato.org/publications/economic-development-bulletin/poor- immigrants-use-public-benefits-lower-rate-poor.Lam, C. & Morse, A. (2012). “U.S. Supreme Court Rules on Arizona’s Immigration Enforcement Law.” 2012. Washington, DC: National Conference of State LegislaturesLeerkes, Arjen, Mark Leach, and James Bachmeier. 2012. “Borders Behind the Border: An Exploration of State-Level Differences in Migration Control and Their Effects on U.S. Migration Patterns.” Journal of Ethnic and Migration Studies 38 (1): 111–29. doi:10.1080/1369183X.2012.640023.Lofstrom, M.; Bohn, S.; & Raphael, S. (2011). “Lessons from the 2007 Legal Arizona Workers Act.” San Francisco, CA: Public Policy Institute of California.Lopez, Mark Hugo, and Susan Minushkin. 2008. “National Survey of Latinos: Hispanics See Their Situation in U.S. Deteriorating; Oppose Key Immigration Enforcement Measures.” Washington, DC: Pew Hispanic Center.Lyles, B. (2015, October 19).Surge in immigrant driver’s licenses may have spurred more organ donors. The Sacramento Bee, Retrieved from https://sacbee.com/.McKinley, J. “Whooping Cough Kills 5 in California; State Declares an Epidemic.” New York Times, June 23, 2010.McWilliams, J. (2009). “Health Consequences of Uninsurance among Adults in the United States.” Milbank Quarterly, 87(2), 443–494. doi:10.1111/j.1468-0009.2009.00564.x Morse, Ann, Lindsay Littlefield, and Leya Speasmaker. 2007. “A Review of State Immigration Legislation in 2005.” Washington, DC: National Conference of State Legislators.Morse, A.; Johnston, A.; Heisel, H.; Carter, A.; Lawrence, M.; & Segreto, J. (2012). “State Omnibus Immigration Legislation and Legal Challenges.” 2012. Denver, CO: National Conference of State Legislatures.National Immigration Law Center. 2014. “Guide to Immigrant Eligibility for Federal Programs Update Page.” www.nilc.org. Accessed October 20. http://www.nilc.org/guideupdate.html.Oritz, E. 2017. “‘Sanctuary’ Cities Targeted by ICE in Immigration Raids as Nearly 500 Arrested.” NBC News Online, https://www.nbcnews.com/storyline/immigration-border-crisis/sanctuary-cities-targeted-ice-immigration- raids-nearly-500-arrested-n805796

16

Parrado, Emilio A. 2012. “Immigration Enforcement Policies, the Economic Recession, and the Size of Local Mexican Immigrant Populations.” The ANNALS of the American Academy of Political and Social Science 641 (1): 16–37. doi:10.1177/0002716211435353.Passel, J. (2019) “Measuring Illegal Immigration: How Pew Research Center Counts Unauthorized Immigrants,” Washington, DC (Pew Hispanic Center). Retrieved April 2020 from https://www.pewresearch.org/fact- tank/2019/07/12/how-pew-research-center-counts-unauthorized-immigrants-in-us/Passel, J., & Cohn, D. (2018) “U.S. Unauthorized Immigrant Total Dips to Lowest Level in a Decade,” Washington, DC (Pew Hispanic Center). Retrieved April 2019 from https://www.pewhispanic.org/2018/11/27/u-s- unauthorized-immigrant-total-dips-to-lowest-level-in-a-decade/.Potochnick, Chen, & Perreira, 2016;Potochnick et al, 2017Potochnick, Stephanie R., and Krista M. Perreira. 2010. “Depression and Anxiety among First-Generation Immigrant Latino Youth: Key Correlates and Implications for Future Research.” The Journal of Nervous and Mental Disease 198 (7): 470–77. doi:10.1097/NMD.0b013e3181e4ce24.Pham, H. (2018). 287(g) Agreements in the Trump Era. Washington and Lee Law Review 75(3): 1253-1286. Retrieved from https://scholarlycommons.law.wlu.edu/cgi/viewcontent.cgi?article=4609&context=wlulr. Precious, T. (2019, May 5). Push for driver’s licenses for undocumented immigrants intensifies at Capitol. The Buffalo News, Retrieved from https://buffalonews.com/.Rosenblum, M. (2015). “Understanding the Potential Impact of Executive Action on Immigration Enforcement.” Washington, DC: Migration Policy Institute.Roth, Benjamin J. 2014. “New Opportunities and Old Barriers: Exploring the Impact of Deferred Action in South Carolina.” In Annual Meeting. Albuquerque, NM.Steege, A.; Baron, S.; Davis S.; Torres-Kilgore, J.; & Sweeney, M. (2009). “Pandemic Influenza and Farmworkers: The Effects of Employment, Social, and Economic factors.” American Journal of Public Health, 99 (Suppl 2), S308– S315. doi:10.2105/AJPH.2009.161091 Urban Institute, “Children of Immigrants Data Tool” (Washington, DC: The Urban Institute, 2019), http://datatool. urban.org/charts/datatool/pages.cfm.Vargas, E. D., & Pirog, M. A. 2016. Mixed-Status Families and WIC Uptake: The Effects of Risk of Deportation on Program Use. Social science quarterly, 97(3), 555-572Waslin M.L. (2013) Driving While Immigrant: Driver’s License Policy and Immigration Enforcement. In: Brotherton D., Stageman D., Leyro S. (eds) Outside Justice. Springer, New York, NYWatson, T. (2010). “Inside the Refrigerator: Immigration Enforcement and Chilling Effects in Medicaid Participation.” Working Paper 16278. National Bureau of Economic Research. http://www.nber.org/papers/w16278.Watson, T. ( 2013). “Enforcement and Immigrant Location Choice.” Working Paper 19626. National Bureau of Economic Research. http://www.nber.org/papers/w19626.Watson (2014)Weathers, A., [ et al]., “Parental Nativity Affects Children’s Health and Access to Care,” Journal of Immigrant and Minority Health / Center for Minority Public Health 10, no. 2 (April 2008): 155–65, doi:10.1007/s10903-007- 9061-y. White, W. [et al] (2014)., “Impact of Alabama’s Immigration Law on Access to Health Care Among Latina Immigrants and Children: Implications for National Reform,” American Journal of Public Health 104, no. 3 : 397–405.Yoshikawa, Hirokazu. 2011. Immigrants Raising Citizens. New York, NY: Russell Sage Foundation.Zimmerman, Wendy, and Karen C. Tumlin. 1999. “Patchwork Policies: State Assistance for Immigrants under Welfare Reform.” Washington, DC: Urban Institute. http://www.urban.org/publications/309007.html.Ziol-Guest, K.M. & Kalil, A. (2012). “Health and Medical Care among the Children of Immigrants,” Child Development 83, no. 5: 1494–1500, doi:10.1111/j.1467-8624.2012.01795.x. 

17

Appendix: Description and history of state-level immigration enforcement policies

Secure Communities and the Priority Enforcement Program: The degree to which states cooperate with federal au-thorities around immigration enforcement revolves largely around local or state responsiveness to the federal Secure Communities program. Secure Communities was launched as a pilot in 2008 under President Bush and expanded nationally during President Obama’s first term. Under Secure Communities, law enforcement agencies submit fingerprints of arrestees for checks against Department of Homeland Security (DHS) databases. Immigra-tion and Customs Enforcement (ICE) officials then decide whether to take enforcement action, including issuing a detainer request to the local law enforcement agency to ask that the individual be held for up to 48 hours so that ICE can take custody. State participation was initially understood to be voluntary, and 35 states signed agreements to participate by 2010. Some state governments also reacted to Secure Communities by passing policies to reject ICE requests in order to preserve trust between immigrant communities and local law enforcement. In 2011, DHS made it clear that the program would operate across the country whether or not states signed on. The program was active in all 50 states and Washington, D.C., by 2012.

Starting in 2011, some states, cities, and counties began pushing back against Secure Communities, passing poli-cies to limit cooperation with ICE detainer requests. Critics argued that the program was leading to the arrest and deportation of noncitizens with minor criminal convictions or arrests but no convictions. This trend coincided with the above move to allow undocumented immigrants to receive driver’s licenses. In late 2014, Secure Communities was replaced with the Priority Enforcement Program (PEP) (Chishti and Hipsman, 2015), which was implemented with narrower enforcement priorities focused on immigrants who were national security threats, had been convict-ed of serious crimes, or were new arrivals in the U.S. PEP also allowed local law enforcement agencies to set further limits on cooperation with federal immigration enforcement (Rosenblum, 2015). President Trump’s January 2017 executive order on interior immigration enforcement ended PEP, reinstated the Secure Communities program, and reestablished expanded enforcement priorities, including undocumented immigrants convicted of any crime, those arrested but not convicted, those who may have committed crimes but have not been arrested, and anyone deemed by an immigration officer to be a public safety threat.

287(g) Agreements: At the same time that local immigration legislation was growing, many communities signed up for a program run by the Department of Homeland Security, forming 287(g) agreements with Immigration and Customs Enforcement (ICE), to allow state or local law enforcement agents to enforce federal immigration law. The 1996 Illegal Immigration Reform and Immigrant Responsibility Act created the 287(g) agreement process, but the first agreement was not signed until 2002 by the Florida Department of Law Enforcement. These agreements allow state or local law enforcement officers to be trained and deputized to carry out certain functions of federal immigration law enforcement, including investigating, apprehending, and detaining noncitizens. Agreements can be signed by city, county, or state law enforcement agencies. The number of 287(g) agreements rose quickly starting in 2005. By 2010, there were 70 active 287(g) agreements with state and local law enforcement agencies (Capps, Rosenblum, Chishti and Rodriguez, 2011). Use of the program fell after 2010, particularly after the Department of Homeland Security (DHS) ended components of the 287(g) program in 2012. In 2016, there were 32 state or local 287(g) agreements, overall. A 2017 executive order by President Trump signaled the beginning of an aggressive marketing program on the part of the federal government and DHS to state and local authorities for again imple-menting all elements of the 287(g) program, which has led to a rapid expansion of the prevalence of these agree-ments (Pham, 2018).

Comprehensive Enforcement State Laws: Some states worked to increase their own authority to address undocu-mented immigration and passed comprehensive immigration enforcement laws. The most widely-known of these was Arizona’s Support Our Law Enforcement and Safe Neighborhoods Act, (introduced as Arizona Senate Bill

18

1070), which aimed to criminalize undocumented status, which is normally a civil, not criminal violation; criminalize seeking work or employment for undocumented immigrants; require local law enforcement to check the immigra-tion status of those who were detained or arrested during the course of normal law enforcement activities; and criminalize sheltering, harboring, or transporting undocumented immigrants. Though most of S.B. 1070’s provisions were eventually struck down by the Supreme Court, others went forward (Lam and Morse, 2012). Between 2010 and 2014, Alabama, Georgia, Indiana, South Carolina, and Utah all passed laws including some of the provisions included in S.B. 1070; all of these also faced legal challenges, and some provisions were never implemented (Morse, Johnston, Heisel, Carter, Lawrence, and Segreto, 2012).

Between 2000-2016, states took diverse stances in their approach to immigration, particularly related to immigra-tion enforcement. In the mid-2000s, local and state governments began passing immigration-related laws at grow-ing rates, such as those prohibiting landlords from renting to undocumented immigrants; restricting or expanding undocumented immigrant youths’ access to in-state tuition and state financial aid; restricting or expanding undocu-mented immigrants’ access to state driver’s licenses and identification documents; requiring local police to enforce federal immigration law; and mandating use of the E-Verify employment eligibility verification system by some or all employers. The number of immigration-related bills enacted by states grew from 37 in 2005 to 208 in 2010 (John-ston and Morse 2011; Morse, Littlefield, and Speasmaker 2007). In the early 2010s, a few states passed legislation that created barriers to civic participation by undocumented immigrants, including criminalization of undocument-ed status. The harshest state enforcement provisions have been struck down by courts or remain embroiled in liti-gation. By 2012, many states and localities started to reduce their participation in federal immigration enforcement efforts through limited cooperation policies, but cooperation with federal immigration enforcement increased after the Trump administration began working in 2017 to increase states’ participation in immigration enforcement.